Claude Dispatch turns your laptop into a remote AI worker

Craig Nash
By
Craig Nash
Tech writer at All Things Geek. Covers artificial intelligence, semiconductors, and computing hardware.
7 Min Read
Claude Dispatch turns your laptop into a remote AI worker

Claude Dispatch is a new feature in Claude Cowork that lets you send tasks from your phone and have them execute automatically on your Mac without touching it again. Anthropic launched Claude Dispatch as a research preview on March 17, 2026, initially available to Max subscribers—the company’s premium tier. The feature represents a fundamental shift: AI moves from real-time chat to asynchronous delegation, turning your idle laptop into a digital coworker.

Key Takeaways

  • Claude Dispatch launched March 17, 2026 as a research preview for Anthropic Max subscribers.
  • Pair your Mac and phone via QR code; tasks execute in a local sandbox with user approval.
  • Supports cross-tool integration: files, browser, code review, document summarization, and more.
  • Requires Mac to remain powered on 24/7; currently Mac-only, not available on Windows or Linux.
  • Orchestration layer auto-selects models based on task complexity, optimizing cost and latency.

How Claude Dispatch Works

Claude Dispatch operates as one persistent conversation running on your Mac desktop. You message Claude from your phone with a task—summarize a document, review code, draft a customer reply—and step away. Claude on your desktop receives the request, routes it to an appropriate model, executes the task in a local sandbox, and returns results in the same thread. No copy-paste. No context-switching between apps. Everything Claude can access on your computer becomes reachable from wherever you are.

Setup requires downloading the Claude Desktop app on Mac and pairing it with your mobile app via QR code scan. Your Mac must remain powered on continuously, as it serves as the execution environment. The system runs tasks in a sandboxed environment and requires your approval before taking actions, ensuring sensitive operations don’t execute without oversight. Files stay local—no data leaves your device.

Real-World Use Cases for Claude Dispatch

The practical applications are where Claude Dispatch stops feeling like a chat feature and starts feeling like infrastructure. You’re reviewing a pull request in Slack on your phone? Send it to Claude. By the time you sit at your desk, the code review is done with suggestions for improvements. You receive a customer message while away? Draft a reply from your phone; Claude composes it on your desktop and waits for approval. You need to turn a spreadsheet into a report? Send the task, return later to polished output.

The orchestration layer behind Claude Dispatch handles model selection automatically. Simple tasks like code linting route to faster, cheaper models. Complex work requiring deeper reasoning—architectural review, strategic analysis—escalates to more capable models like Claude Opus or Sonnet. This intelligent routing reduces latency and cost without forcing you to choose the right model yourself. The system also handles failures gracefully, retrying or escalating as needed.

Setup and Limitations of Claude Dispatch

Getting started is straightforward but has real constraints. You need an Anthropic Max subscription—the company’s premium tier—and a Mac running Claude Desktop. Windows and Linux users are out. Your Mac must stay on 24/7; if it sleeps or shuts down, tasks queue until it returns. This architecture trades convenience for privacy: everything executes locally, no cloud processing of your sensitive work.

The permission model is critical. When you pair your phone to your Mac, Claude gains access to everything your desktop can access—files, browser history, applications, tools. For routine tasks like code review or document summarization, this is fine. For sensitive work involving financial data, personal information, or proprietary systems, you should only enable Claude Dispatch on a machine you’re comfortable with worst-case access on. This is not a limitation of the feature; it’s a reality of any remote delegation system.

Claude Dispatch vs. Traditional AI Chat

Traditional AI chat is synchronous. You type, wait for an answer, copy the result, paste it elsewhere. Claude Dispatch is asynchronous. You send work, step away, retrieve it later in context. This eliminates friction in fragmented workflows—the reality of modern work, where you’re jumping between your phone, email, Slack, and your laptop. Instead of holding context in your head or switching apps repeatedly, Claude Dispatch keeps everything in one persistent thread.

Compared to general AI services like ChatGPT, Claude Dispatch operates entirely locally. Your data never leaves your machine. Compared to traditional automation tools, it requires no configuration of workflows, rules, or integrations. You simply describe what you need in natural language, and Claude figures out which tools to use and in what order. This represents a meaningful step toward AI systems that function more like operating systems than chatbots.

Is Claude Dispatch ready for everyday use?

Claude Dispatch is a research preview, not a finished product. Anthropic is actively refining it. The feature works well for the demonstrated use cases—code review, document processing, message drafting—but broad user testing data doesn’t exist yet. If you’re an Anthropic Max subscriber with a Mac you’re willing to keep powered on, it’s worth trying. If you’re waiting for Windows support or lower-cost tiers, you’ll be waiting longer.

What does Claude Dispatch cost?

Claude Dispatch requires an Anthropic Max subscription, which is the company’s premium tier. Anthropic has not published specific pricing in the research preview materials. The feature itself has no separate cost beyond the Max subscription, though using it will consume tokens based on task complexity and the models selected.

Can I use Claude Dispatch on Windows or Linux?

Not currently. Claude Dispatch requires Claude Desktop on Mac and is not available for Windows or Linux. Anthropic has not announced plans to expand platform support.

Claude Dispatch signals where AI is heading: away from chat interfaces and toward persistent, context-aware delegation. You’re not talking to an AI anymore. You’re directing one to work while you do something else. For knowledge workers juggling multiple contexts, that’s genuinely useful—if you can afford the subscription and keep your Mac running.

Edited by the All Things Geek team.

Source: Tom's Guide

Share This Article
Tech writer at All Things Geek. Covers artificial intelligence, semiconductors, and computing hardware.